The investigators propose to obtain longitudinal follow up data and long-term (5 year) outcomes on the HCSUS sample using face to face interviews and searches of appropriate national death indexes. They also propose additional analyses of previously collected data after linking it with the follow-up data. All follow-up data will be derived from the patient survey, no additional chart extraction or provider interviews are planned. The budget is approximately $2.6 million over 2 years and represents an 18% increase over the budget of the last submission. The investigators offer two other options: 1. Telephone interview only with a budget of 2.2 million over two years (same as prior budget) 2. Face-to-face interviews and an additional blood draw with central determination of virologic and immunologic status and specimen banking with a budget of 3.5 million over two years (60% increase over previous submission). Of note, if the interviews were excluded from this proposal, the budget would be substantially reduced. The budget for the NORC subcontract alone is $874,207. The specific aims for this 5-year follow up data [all patient reported save mortality] are: 1. Measure patterns and processes of healthcare and other services including: use, cost, barriers to use, case-management, quality of care, and health behaviors. 2. Measure clinical and policy relevant outcomes including: mortality, CD4 count, symptoms, complications of HIV or its treatment, quality of life, disability, and work status. 3. Identify important predictors of long-term clinical and policy related outcomes. Candidate predictors include: health status, treatment, insurance, case management, unmet needs, provider knowledge, and other variables. The revised proposal has the following hypotheses: 1. The gap in access to antiretroviral therapy and service use between disadvantaged and advantaged groups will persist over time, be explained by greater barriers to receipt of care, and predict differences in survival, and in clinical and quality of life outcomes. 2. Insurance status and generosity of public insurance as measured by state policies will predict clinical outcomes, employment status and other labor market outcomes, subsequent insurance status, and use and cost of efficacious chemotherapy, hospital, and other care. 3. Clinical status, including virologic and immunologic parameters, therapy and adherence to it, opportunistic HIV complications, chronic non-HIV-related comorbidities, and functional/symptom status will predict survival, disease progression, quality of life, and use of care better than conventional staging. 4. Time on adequate ARV therapy and overall quality, as measured by an index of indicators, will predict both mortality and other outcomes. 5. Provider specialty training and experience, and practice site characteristics will predict mortality and health related quality of life through its effects on prescribing, adherence, and other care processes.